Abstract

Closed-loop combustion control on internal combustion engines has historically been seen as an important way to improve fuel economy and reduce CO2 emissions, but also to reduce harmful tailpipe emissions. More recently, it is viewed as an important part of achieving stable operation of IC engines running on zero carbon fuels. Combustion control requires knowledge of the instantaneous cylinder pressures, which often show considerable cycle-by-cycle variability owing to the stochastic nature of combustion. Direct measurement of cylinder pressure is achievable using in-cylinder pressure sensors, fitted routinely for engine development and test programmes, or in high-value motor sport. Commercial pressure sensors are however expensive, and are generally designed to endure the very hostile in-cylinder conditions for only relatively limited duration without need of servicing and recalibration. Direct sensing, therefore, involving commercial pressure sensors on production engines, is still problematic. Effort has been devoted over the past 25 years on finding a suitable indirect way of reconstructing engine cylinder pressure. This talk will discuss the use of inverse modelling techniques that have been adopted to exploit engine crank dynamics, block vibrations, and acoustic emissions, all of which can be measured with relatively inexpensive sensors.